CBRシステムの構築環境

概要

論文の詳細を見る
Many kinds of CBR (Case-Based Reasoning) technologies have been developed and applied to a broad range of problems. However, there are no consistent methodologies nor support tools to develop CBR systems. This paper proposes a classification of problem types and a CBR system building support environment. The basic idea of the classification is that case functions are crucial to CBR problem solving. The case functions determine the problem solving process which can be represented by a specific set of generic tasks. The generic tasks are the units of human experts' thinking processes and are used as the functional elements of systems. We identified 6 CBR problem types (Selection, Modification, Assembly, Mediation, Estimation and Complement) and 10 generic tasks. Each generic task is implemented into several building blocks. The proposed CBR system building environment consists of a library of the building blocks, strategic knowledge and a case base manager. The main feature of the environment is the support for constructing nested architectures, that is, recursive, built-in or repetitive models, which, we insist, are the architectural feature of CBR systems. Finally, several systems are built using the environment, and its effectiveness is shown.
社団法人人工知能学会の論文
1994-01-01